What happened
Anthropic confirmed Wednesday it hired Chad Jones, the Stanford GSB economist whose 1995 paper on R&D and growth is still a fixture of graduate macro reading lists, to work on the economics of catastrophic AI risk. Crypto Briefing first reported the move at 18:07 UTC. Jones is not joining the alignment team or the policy team in the conventional sense.
He's there to build the models that price the trade-off between AI-driven productivity gains and tail-risk losses, the kind of framing that ends up in congressional testimony and OECD working papers. Anthropic has not published a research agenda for Jones, and the company declined to specify whether he joins full-time or in an advisory capacity. No compensation figures, no equity terms, no start date beyond "this summer" were disclosed.
Why it matters
AI safety arguments have lived in two camps for three years. The technical camp talks about interpretability, jailbreaks, and capability evals. The doom camp talks about probabilities that don't sit in any spreadsheet.
Jones is a bridge. His prior work on idea production and the "are ideas getting harder to find" debate is exactly the toolkit you'd use to model what happens to growth, employment, and welfare if frontier models keep scaling. That matters now because the US AI Safety Institute, the UK AISI, and the EU AI Office are all writing rules that need a number attached to "catastrophic risk" before they bind anything.
Anthropic, which has been the most vocal frontier lab calling for compute-threshold regulation, gets a credentialed economist to put a price tag on its policy asks. The signal to rivals is harder: OpenAI and Google DeepMind don't have a comparable hire on the economics side, and they'll either match it or cede the framing.